use "legacy" dataclass at operator level and separate TestNetOutput from TrainNetOutput

Summary: Pull Request resolved: #444

Differential Revision: D41828774

fbshipit-source-id: 833dea0e79eccdb8396bafb6c73f5255fa3cfddb

Author: Yanghan Wang

d2go/trainer/api.py

@@ -11,14 +11,20 @@
 
 from d2go.evaluation.api import AccuracyDict, MetricsDict
 
-# TODO (T127368935) Split to TrainNetOutput and TestNetOutput
+
 @dataclass
 class TrainNetOutput:
     accuracy: AccuracyDict[float]
     metrics: MetricsDict[float]
-    # Optional, because we use None to distinguish "not used" from
-    # empty model configs. With T127368935, this should be reverted to dict.
-    model_configs: Optional[Dict[str, str]]
+    model_configs: Dict[str, str]
+    # TODO (T127368603): decide if `tensorboard_log_dir` should be part of output
+    tensorboard_log_dir: Optional[str] = None
+
+
+@dataclass
+class TestNetOutput:
+    accuracy: AccuracyDict[float]
+    metrics: MetricsDict[float]
     # TODO (T127368603): decide if `tensorboard_log_dir` should be part of output
     tensorboard_log_dir: Optional[str] = None
 

tools/lightning_train_net.py

@@ -12,7 +12,7 @@
 from d2go.runner.callbacks.quantization import QuantizationAwareTraining
 from d2go.runner.lightning_task import DefaultTask
 from d2go.setup import basic_argument_parser, prepare_for_launch, setup_after_launch
-from d2go.trainer.api import TrainNetOutput
+from d2go.trainer.api import TestNetOutput, TrainNetOutput
 from d2go.trainer.helper import parse_precision_from_string
 from d2go.trainer.lightning.training_loop import _do_test, _do_train
 from detectron2.utils.file_io import PathManager
@@ -103,7 +103,7 @@ def main(
     output_dir: str,
     runner_class: Union[str, Type[DefaultTask]],
     eval_only: bool = False,
-) -> TrainNetOutput:
+) -> Union[TrainNetOutput, TestNetOutput]:
     """Main function for launching a training with lightning trainer
     Args:
         cfg: D2go config node
@@ -123,18 +123,22 @@ def main(
         logger.info(f"Resuming training from checkpoint: {last_checkpoint}.")
 
     trainer = pl.Trainer(**trainer_params)
-    model_configs = None
+
     if eval_only:
         _do_test(trainer, task)
+        return TestNetOutput(
+            tensorboard_log_dir=trainer_params["logger"].log_dir,
+            accuracy=task.eval_res,
+            metrics=task.eval_res,
+        )
     else:
         model_configs = _do_train(cfg, trainer, task)
-
-    return TrainNetOutput(
-        tensorboard_log_dir=trainer_params["logger"].log_dir,
-        accuracy=task.eval_res,
-        metrics=task.eval_res,
-        model_configs=model_configs,
-    )
+        return TrainNetOutput(
+            tensorboard_log_dir=trainer_params["logger"].log_dir,
+            accuracy=task.eval_res,
+            metrics=task.eval_res,
+            model_configs=model_configs,
+        )
 
 
 def argument_parser():

tools/train_net.py

@@ -9,7 +9,6 @@
 import sys
 from typing import List, Type, Union
 
-import detectron2.utils.comm as comm
 from d2go.config import CfgNode
 from d2go.distributed import launch
 from d2go.runner import BaseRunner
@@ -22,7 +21,7 @@
     setup_before_launch,
     setup_root_logger,
 )
-from d2go.trainer.api import TrainNetOutput
+from d2go.trainer.api import TestNetOutput, TrainNetOutput
 from d2go.trainer.fsdp import create_ddp_model_with_sharding
 from d2go.utils.misc import (
     dump_trained_model_configs,
@@ -40,7 +39,7 @@ def main(
     runner_class: Union[str, Type[BaseRunner]],
     eval_only: bool = False,
     resume: bool = True,  # NOTE: always enable resume when running on cluster
-) -> TrainNetOutput:
+) -> Union[TrainNetOutput, TestNetOutput]:
     runner = setup_after_launch(cfg, output_dir, runner_class)
 
     model = runner.build_model(cfg)
@@ -58,9 +57,8 @@ def main(
         model.eval()
         metrics = runner.do_test(cfg, model, train_iter=train_iter)
         print_metrics_table(metrics)
-        return TrainNetOutput(
+        return TestNetOutput(
             accuracy=metrics,
-            model_configs={},
             metrics=metrics,
         )